EUROPEAN JOURNAL OF ENTOMOLOGYENTOMOLOGY ISSN (online): 1802-8829 Eur. J. Entomol. 113: 531–536, 2016 http://www.eje.cz doi: 10.14411/eje.2016.072 ORIGINAL ARTICLE

Comparative effi ciency of traps for horse fl y (Diptera: Tabanidae) survey in riparian oak-ash forests in Danube fl oodplain

ALMA MIKUŠKA1, SELMA MLINARIĆ1, LIDIJA BEGOVIĆ1 and ERIN CURRAN 2

1 Department of Biology, Josip Juraj Strossmayer University of Osijek, Ulica cara Hadrijana 8/A, HR-31000 Osijek, Croatia; e-mails: [email protected], [email protected], [email protected] 2 Computer and Information Sciences Department, University of St. Thomas, 2115 Summit Ave., St. Paul, MN 55105, USA; e-mail: [email protected]

Key words. Diptera, Tabanidae, horse fl ies, Box trap, Nzi trap, Canopy trap, catching effi ciency

Abstract. The objective of this work was to evaluate the relative trap effi ciency of the three modifi ed traps baited with 1-octen-3-ol as attractant, Box, Nzi and Canopy, for horse fl y (Tabanidae) survey. Nine traps (three traps per trap type) were tested in 3 × 3 Latin square designs during summer 2011 in riparian oak-ash forests in the Danube fl oodplain of Croatia. Overall, the traps caught 1,986 specimens of 11 horse fl ies species during the study period. The most abundant species caught was bromius (58%), followed by Tabanus tergestinus (21%), Tabanus maculicornis (16%), Tabanus sudeticus (2%), Atylotus loewianus (1%) and Tabanus autumnalis (1%). Both the Box traps and the Nzi traps had signifi cantly greater catch effi ciencies than the Canopy 2 2 traps (βbox = 2.47, X = 65.48, df = 1, P < 0.001 and βnzi = 1.54, X = 25.12, df = 1, P < 0.001, respectively). Nzi traps had a catch rate 4.65 (95% CI: 2.55, 8.48) times greater than Canopy traps and Box traps had a catch rate 11.77 (95% CI: 6.48, 21.39) times greater than Canopy traps. Based on Shannon-Wiener diversity indices Nzi traps were better suited for diversity survey and had higher (H’ = 1.33) indices than Box (H’ = 1.08) or Canopy traps (H’ = 1.00).

INTRODUCTION & Hogsette, 1994). Trapping devices for adult horse fl ies Horse fl ies (Tabanidae) are part of the Dipteran family are based on fl ight interception or attraction using a visu- comprising over 4,400 species known worldwide (Rosk- al stimulus and/or a natural or synthetic odour. Recently, ov et al., 2015). The presence of water, hosts, and forests traps based on the attraction of horse fl ies to horizontally- are primary and required conditions for the occurrence of polarized light have been developed (Horváth et al., 2008; horse fl ies (Chvála et al., 1997). In dry areas and steppes Blahó et al., 2012). Odour-baited sticky boards are traps without trees, or even along large rivers that are not fringed based on attraction (Hall et al., 1998) and traps with elec- with forests and other vegetation, horse fl ies are mostly ab- tric grids are based on fl ight interception (Muzari et al., sent (Chvála et al., 1997). However, under specifi c condi- 2010) or may be associated with a visual or odour attractant tions they can occur in larger numbers in semi-arid steppes (Krčmar et al., 2014). An optimal survey trap for all horse (Baldacchino et al., 2013). fl ies species is diffi cult to achieve (Gibson & Torr, 1999) Host-seeking females are considered to be major pests of because different trap types may capture different numbers ungulates world-wide. Female horse fl ies feed on domestic of individuals and species (Mihok et al., 2006; Krčmar & and wild (Baldacchino et al., 2014b) and are me- Rupnik Poklukar, 2011). Very few studies measuring trap chanical vectors of different pathogens, such us the equine effi ciency with horse fl ies have been completed (Mihok infectious anemia virus and Besnoitia besnoiti, between et al., 2006; Mihok & Carslon 2007; Van Hennekeler et the animals (Desquesnes et al., 2009; Votýpka et al., 2015). al., 2008; Krčmar et al., 2014). Nzi, Box and Canopy traps Horse fl ies are generally diffi cult to control and currently have been used for catching horse fl ies in different fi eld there are no adequate control measures for reducing their studies (Krčmar et al., 2006, 2010, 2014; Baldacchino et populations (Barros & Foil, 2007). Traps are designed to al., 2013), but a comparison and evaluation of trap effi - attract targeted using sensory cues and they are ciency has never been tested. practical tools for vector surveillance. However, traps have The present study aimed to evaluate the relative trapping been only occasionally tested for control measures (Balda- effi ciency of modifi ed designs of Nzi, Box and Canopy cchino et al., 2014a). Control of horse fl ies must be based traps for horse fl y survey. Trapping effi ciency was defi ned on protective actions against adults, rather than larvae (Foil as the number of horse fl ies caught in a given time period

Final formatted article © Institute of Entomology, Biology Centre, Czech Academy of Sciences, České Budějovice. An Open Access article distributed under the Creative Commons (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

531 Mikuška et al., Eur. J. Entomol. 113: 531–536, 2016 doi: 10.14411/eje.2016.072

Fig. 1. Modifi ed Box (A), Nzi (B) and Canopy (C) at study site.

(trap/day). The aim of this study was to evaluate the im- The tent-like Canopy traps were constructed according to the pact of trap design and construction on trapping effi ciency, design of Hribar et al. (1991). Four-sided pyramidal structures, rather than to evaluate horse fl y attraction to visual stimuli 250 cm tall with 120 cm wide openings, were made from black such as trap colour. and white cotton fabric. The collecting caps were made from polyester mosquito mesh, 20 cm wide and 35 cm high, and were MATERIALS AND METHODS placed on the top of the trap (Fig. 1C). The entrance to the Cano- py trap was 50 cm from the ground. Study site and traps To improve collection performance of the traps, 1-octen-3-ol The study was carried out in riparian oak-ash forests in the (98% pure; Sigma-Aldrich Chemie GmbH, Germany) was used Danube fl oodplain of Croatia (45°41´55.59˝S, 18°49´20.74˝E) in as an attractant. Plastic vials with 4 mL of 1-octen-3-ol and 5 cm Europe. The area has a typical continental climate with wide an- long cotton were attached to wooden sticks inside all of the traps nual fl uctuations of air temperatures and precipitation, as well as and placed 30 cm below the top of traps. The attractant evapora- four distinctive seasons. The average annual air temperature is tion was measured at the study site over 24 h (with a daily mean 10.5°C (with a maximum temperature of 39°C in July). The aver- temperature of 20.4°C and relative moisture of 88% without the age annual precipitation is 687 mm with a peak during May–June wind) to test the possibility that 1-octen-3-ol would evaporate dif- (data were provided by the Croatian Meteorological and Hydro- ferently, depending on trap type. The evaporation of the attractant logical Service). Ecological studies have shown that horse fl ies was up to 1 mL/day in every trap type. Evaporation of 1-octen- are most active at pasture-forest ecotone near their developmen- 3-ol in Box traps was faster than in Nzi or Canopy traps. Conse- tal, host-seeking and resting sites (Baldacchino et al., 2013) and quently, 1-octen-3-ol was added daily, up to the amount of 4 mL, so traps were placed on the pasture at 60 m distance and parallel to ensure that there was a suffi cient supply of attractant during to the forest edge. throughout the trapping time. Modifi ed Box traps consisted of 80 × 80 × 60 cm four-sided Study design plywood boxes, open underneath with a metal screen roof as described by Hansens & Race (2011). The opening of the trap A total of 24 sampling days, eight continuous days per month, was set 50 cm above the ground. The collectors on these traps were conducted between June and August 2011. The trapping pe- were modifi ed and a 20 cm wide opening was cut in the center riod was between 8:00 am and 8:00 pm each day. The experiment of the screen roof, rather than in the cones. Removable collecting consisted of three 3 × 3 Latin square designs. Nine traps (includ- caps made from polyester mosquito mesh (the same size as those ing three traps of each type) were placed 50 m apart from one on canopy and Nzi traps) were used. The Box traps were painted another. The positions of the traps were exchanged every month outside with a violet paint while the inner side of the box was left to nullify any possible effect of trap position on trap performance. unpainted (Fig. 1A). The painting of the traps and the quantitative Thus, each trap type was located at each of nine possible posi- assessment of colours was performed according to the methods tions during the course of this experiment. described in Krčmar et al. (2014). An area of approximately 5 m in diameter was cleared of grass Modifi ed Nzi traps were 1 m cloth triangular traps constructed around each trap. The collection caps were emptied every morn- as described by Mihok (2007). The traps were made from black ing, and collected horse fl ies were preserved in 70% ethanol. The and blue cotton/polyester (65/35%) fabric (Tex Ingro S.R.L. identifi cation and the nomenclature followed Chvála et al. (1972) Italy). The blue fabric had peak wavelengths at 491 nm, as recom- and Chvála (1988). mended by Mihok (2007). The modifi cation to the original design Statistical analysis included an opening, 20 cm wide, at the top of the trap. A remov- A negative binomial regression model incorporating an overd- able collecting cap, the same size as those on canopy traps and ispersion parameter estimate of 1.0 was used to predict horse fl y made from polyester mosquito mesh, was also used. The traps catch effi ciency from the type of trap used. Negative binomial re- were set on the ground with the entrance facing the forest edge gression is an appropriate analytic strategy given the strong right (Fig. 1B). skew observed in the dependent variable and the catch effi ciency

532 Mikuška et al., Eur. J. Entomol. 113: 531–536, 2016 doi: 10.14411/eje.2016.072 differences attributable to trap type over a defi ned period of data collection. To create a model of catch effi ciency, the variable “trap type” was dummy coded to generate two binary predictors (Box trap and Nzi trap); canopy trap was used as the reference category. The log of the outcome variable, “total horse fl y count”, an indicator of catch effi ciency, was modeled as a function of the two identifi ed binary predictors and the overdispersion parameter estimate. Models were created for the overall results (utilizing data from all horse fl y specimens collected) and separately for the T. bromius, T. tergestinus, and T. maculicornis species of the horse fl y. Other horse fl y species were present in numbers too low for such modelling. In addition to overall tests of model signifi cance, several ad- ditional statistical results were evaluated. Regression model coeffi cients (β) were examined to understand the nature of the association between the predictor and the outcome variable. Con- tributions of each predictor to the model were evaluated for sta- 2 tistical signifi cance using Wald X test statistics. Incidence rate Fig. 2. Trap effi ciencies (median number of horse fl ies/trap days) ratios and their 95% confi dence interval limits were calculated for three trap types. Numbers above the circles represent outliers, to quantify differences between trap types in catch rates. Finally, boxes represent upper and lower quartile, while whiskers repre- pairwise comparisons of trap types on catch effi ciencies were sent minimum and maximum values except outliers. calculated to identify statistically signifi cant differences between trap types regarding catch effi ciency. A Bonferroni adjustment The modifi ed Box traps consistently caught more horse was applied to the pairwise comparison signifi cance values to enhance protection against Type I error. All statistical analyses fl ies than either modifi ed Nzi or Canopy traps (Fig. 2). The were performed using IBM SPSS Statistics for Windows, Ver. overall model predicting total horse fl y count from predic- 2 20.0 (IBM Corporation, Armonk, NY). tors Box trap and Nzi trap is statistically signifi cant, X Data on horse fl y fauna from a previous study were used to = 57.54, df = 2, P < 0.001. This result indicates that this assess the effi ciency of different trap types for surveillance model represents a statistically signifi cant improvement in (Mikuška, 2010). During this survey, carried out from 2004– fi t over a null model containing zero predictors. Param- 2009, 23 species with 12,452 individuals in total were caught by eter estimate statistics suggest that both the Box traps and Canopy traps baited with different natural and synthetic attract- the Nzi traps have signifi cantly greater catch effi ciencies ants. Different trap types were compared in terms of effi ciency than the Canopy traps (β = 2.47, X 2 = 65.48, df = 1, P < for surveillance using Shannon-Wiener index of diversity (Krebs, box 0.001 and β = 1.54, X 2 = 25.12, df = 1, P < 0.001, respec- 1989). nzi tively). Incidence rate ratios suggest that for all species of RESULTS horse fl y included in this study, Nzi traps are expected to In total, 1,986 horse fl y specimens were collected during have a catch rate 4.65 (95% CI: 2.55, 8.48) times higher the sampling period. Out of 11 collected species, the most than Canopy traps. The Box traps are expected to have a abundant was T. bromius with 58% of the total catch, fol- catch rate 11.77 (95% CI: 6.48, 21.39) times greater than lowed by T. tergestinus (21%) and T. maculicornis (16%). Canopy traps, holding all other variables in the model con- All other species each represented less than 2% of the total stant. Pairwise comparisons indicate that box traps (x̅ = catch (Table 1). The same catching order of three most 55.92) demonstrate better catch effi ciency over both Nzi abundant species was maintained among the three trap traps (x̅ = 22.08; P = 0.019) and Canopy traps (x̅ = 4.75; types (Table 1). Modifi ed Nzi traps were more effi cient P < 0.001). Moreover, Nzi traps demonstrate signifi cantly in catching T. sudeticus and T. autumnalis than Box and greater catch effi ciency than Canopy traps (P = 0.001). Canopy traps (Table 1).

Table 1. Number of species and specimens caught per each trap type during the study period. Modifi ed Box Modifi ed Nzi Modifi ed Canopy Total Species / Trap type No: % No: % No: % No: % L.,1758 817 61 254 48 80 70 1151 58 Tabanus tergestinus Egger, 1859 241 18 154 29 19 17 414 21 Tabanus maculicornis Zetterstrdt, 1842 238 18 65 12 8 7 311 16 Tabanus sudeticus Zeller, 1842 5 0.4 35 7 – 40 2 Atylotus loewianus (Villeneuve, 1920) 26 2 – 2 2 28 1 Tabanus autumnalis L., 1761 10 0.8 15 3 1 0.9 26 1 Haematopota pluvialis (L., 1758) 2 0.2 1 0.2 2 1.8 5 0.3 Hybomitra muehlfeldi (Brauer, 1880) 1 0.07 3 0.6 – 4 0.2 Haematopota italica Meigen, 1804 1 0.07 1 0.2 1 0.9 3 0.2 Hybomitra ciureai (Séguy, 1937) 1 0.07 1 0.2 1 0.9 3 0.2 L., 1758 – 1 0.2 – 1 0.05 Total 1342 530 114 1986

533 Mikuška et al., Eur. J. Entomol. 113: 531–536, 2016 doi: 10.14411/eje.2016.072

Table 2. Capture rates of three trap types for each species (mean ± SD). Mean number of each species caught per trap (mean ± SD) Species Modifi ed Box trap Modifi ed Nzi trap Modifi ed Canopy trap T. bromius 11 (± 13) 4 (± 8) 1 (± 3) T. tergestinus 3 (± 10) 2 (± 6) 0.3 (± 0.8) T. maculicornis 3 (± 10) 0.9 (± 3) 0.1 (± 0.5) A. loewianus 0.4 (± 1) – 0.03 (± 0.2) T. autumnalis 0.1 (± 0.4) 0.2 (± 0.6) 0.01 (± 0.1) T. sudeticus 0.1 (± 0.3) 0.5 (± 1.4) – H. pluvialis 0.02 (± 0.2) 0.01 (± 0.1) 0.02 (± 0.2) H. italica 0.01 (± 0.1) 0.01 (± 0.1) 0.01 (± 0.1) H. ciureai 0.01 (± 0.1) 0.01 (± 0.1) 0.01 (± 0.1) H. muehfl eldi 0.01 (± 0.1) 0.04 (± 0.3) – T. bovinus – 0.01 (± 0.1) – Total 18 (± 39) 7 (± 16) 2 (± 4)

Capture effi ciencies of three trap types for each species are comparisons indicate that Box traps (x̅ = 10.04) and Nzi presented in Table 2. traps (x̅ = 6.42) showed statistically similar catch effi cien- Evaluation of traps for surveillance cies (P = 0.475). However, both Box traps and Nzi traps demonstrated signifi cantly greater catch effi ciencies than The fauna of the studied area consists 23 horse fl y spe- Canopy traps (x̅ = 0.79; P < 0.001 and P < 0.001, respec- cies recorded during the 2004–2009 period (Mikuška, tively). 2010). During this study period traps have caught only 11 (c) Tabanus maculicornis: The model predicting total T. species in total (Box and Nzi – 10 species, Canopy – 8 maculicornis horse fl y count from predictors Box trap and species) (Table 1). However, traps caught the nine most Nzi trap is statistically signifi cant, X 2 = 69.99, df = 2, P abundant species in the area. Comparing Shannon-Wiener < 0.001. For this species, pairwise comparisons suggested diversity indices revealed that Nzi traps had higher (H’ = that Box traps (x̅ = 9.92) showed signifi cantly greater catch 1.33) indices than Box (H’ = 1.08) or Canopy traps (H’ = effi ciency than both Nzi traps (x̅ = 2.71; P = 0.004) and 1.00), but lower Shannon-Wiener diversity index than for Canopy traps (x̅ = 3.33; P < 0.001). Additionally, Nzi traps the area (H’ = 1.50). demonstrated signifi cantly greater catch effi ciency than Species-specifi c results canopy traps (P = 0.001). Capture effi ciencies and catch rates for three most abun- DISCUSSION dant horse fl y species are presented in Table 3. (a) Tabanus bromius: Similar to the overall results, the Our study demonstrated that modifi ed Box traps had model predicting total T. bromius horse fl y count from pre- higher trapping effi ciency than both modifi ed Nzi and dictors Box trap and Nzi trap is statistically signifi cant, X 2 Canopy traps using 1-octen-3-ol as attractant for the three = 53.71, df = 2, P < 0.001. Pairwise comparisons indicate traps. This result may be attributed to two main features that Box traps (x̅ = 34.04) possess superior catch effi ciency of trap construction and design: trap opening size/position over both Nzi traps (x̅ = 10.58; P = 0.005) and Canopy and attractant release rate. Box traps have a 1 m2 opening traps (x̅ = 3.33; P < 0.001); Nzi traps also demonstrate sig- on the bottom of the trap and the trap is placed 50 cm above nifi cantly greater catch effi ciency over Canopy traps (P = the ground; this large, accessible opening allows horse fl ies 0.007) for this species. to enter the trap from all sides. By contrast, Nzi traps have (b) Tabanus tergestinus: The model predicting total T. an opening that is only half of the size (0.5 m2) of the Box tergestinus horse fl y count from predictors Box trap and trap opening which may limit accessibility. Furthermore, Nzi trap is statistically signifi cant (X 2 = 46.89, df = 2, P because these traps are set on the ground, the Nzi trap < 0.001). For this particular species of horse fl y, pairwise opening serves to intercept the fl ight of horse fl ies com-

Table 3. Catch effi ciencies (β) and catch rates for three most abundant horse fl y species. Species / Trap type Tabanus bromius Tabanus tergestinus Tabanus maculicornis 2.32 2.54 3.39 Box χ2 = 55.63; df = 1 χ2 = 46.06; df = 1 χ2 = 54.16; df = 1 P < 0.001 P < 0.001 P < 0.001 Catch effi ciency (β) 1.16 2.09 2.10 Nzi χ2 = 13.38; df = 1 χ2 = 30.74; df = 1 χ2 = 19.62; df = 1 P < 0.001 P < 0.001 P < 0.001 10.21 12.68 29.75 Box Catch rate vs. 95% CI: 5.55, 18.81 95% CI: 6.09, 26.42 95% CI: 12.05, 73.44 Canopy trap 3.18 8.11 8.13 Nzi 95% CI: 1.71, 5.90 95% CI: 3.87, 16.98 95% CI: 3.22, 20.53 df – degrees of freedom, CI – confi dence interval 534 Mikuška et al., Eur. J. Entomol. 113: 531–536, 2016 doi: 10.14411/eje.2016.072 ing from only one direction. Like Box traps, Canopy traps that Nzi traps were better suited for diversity survey than also have a large opening on the bottom (120 cm2), but it either Box traps or Canopy traps. This result could be re- is diamond shaped when the trap is erected. Furthermore, lated to the aspects of Nzi trap design that intercept the the roof of a Box trap is made of a metal insects screen that straight fl ight of the horse fl y, while the Box and Canopy provides a 1 m2 large opening where attractant can freely traps require that the horse fl y enter the trap from below. disperse into the surrounding area. By contrast, Canopy Moreover, different horse fl y species have unique feed- traps have a very small (up to 10 cm2) opening on top and ing behaviours and patterns (Krčmar & Galić, 2004). Box most of the evaporated attractant is held within the tent- traps, given their particular design and construction, may like structure. Indeed, in this study, Box traps had a faster be better suited for catching species that feed on the ventral rate of attractant evaporation than Nzi or Canopy traps (but side of the prey. not higher than 1 mL/day). Higher evaporation rates could Control of insect populations, particularly those that are lead to higher numbers of horse fl ies attracted to the trap considered pests, is generally diffi cult concerning the re- and consequently to higher catch rates. duction of their populations (Barros & Foil, 2007). Horse In addition to yielding higher catch rates, the robust con- fl ies are not the exception, and their control must be based struction of modifi ed Box traps made them highly practical on protective measures against adults (Foil & Hogsette, for long-term horse fl y survey and monitoring. The fabric 1994). While it is evident that traps can not guarantee the of Nzi traps was occasionally damaged by grasshoppers overall population decline of any horse fl y species, their (Orthoptera: Acrididae) that were chewing holes in the use is recommended as a tool to eliminate excessive num- netting of the traps. Constant maintenance was required bers of pests locally and thus reduce the impact of pests on to keep the Nzi traps operational and made their routine domestic animals (Baldacchino et al., 2014a). Our study use in the fi eld rather diffi cult. This assertion is consist- showed that Box traps are best suited for such tasks con- ent with the fi ndings of Taylor & Berkebile (2006). On the cerning the numbers of horse fl ies caught and their rela- other hand, the modifi ed Nzi traps were more stable than tive trapping effi ciency. This result was consistent regard- the modifi ed Canopy traps. Morevoer, the combination ing both the overall number of captured horse fl ies and the of blue paint and octenol as an attractant may have made most abundant species in the area. modifi ed Nzi traps more effi cient than the modifi ed black ACKNOWLEDGEMENTS. This study was supported by the Min- and white Canopy traps baited with the same attractant. istry of Science, Education and Sport of the Republic Croatia (re- The relatively weak performance of the Canopy traps in search grant number: 285-0731674-2200). Authors wish to thank the present study is consistent with the results of previous S. Krčmar for his help in this research. studies in North America and Australia (Mihok et al., 2006; Van Hennekeler et al., 2008). REFERENCES

Trap design and construction are not likely the only BALDACCHINO F., CADIER J., PORCIANI A., BUATOIS B., DORMONT L. factors impacting the effi ciency of the traps. The effects & JAY-ROBERT P. 2013: Behavioural and electrophysiological of specifi c materials and dye colors (Mihok & Carlson, responses of females of two species of tabanid to volatiles in 2007), as well as different natural and synthetic attractants urine of different mammals. — Med. Vet. Entomol. 27: 77–85. (Krčmar, 2005; Krčmar et al., 2006, 2010), have been stud- BALDACCHINO F., DESQUESNES M., MIHOK S., FOIL L.D., DUVALLET ied. According to Egri et al. (2012), linearly polarized light G. & JITTAPALAPONG S. 2014a: Tabanids: Neglected subjects of attracts host-seeking females, but sensitivity for polariza- research, but important vectors of disease agents! — Infect. Genet. Evol. 28: 596–615. tion of light has not been studied at the structural level of BALDACCHINO F., GARDÈS L., DE STORDEU R E., JAY-ROBERT P. & the horse fl y eyes and may differ among species (Baldac- GARROS C. 2014b: Blood-feeding patterns of horse fl ies in the chino et al., 2014a). French Pyrenees. — Vet. Parasitol. 199: 283–288. Traps are commonly used devices for monitoring, sur- BARROS A. & FOIL L. 2007: The infl uenc e of distance on move- veillance, and control of insect populations. Monitor- ment of tabanids (Diptera: Tabanidae) between horses. — Vet. ing programs may have goals that are different from the Parasitol. 144: 380–384. evaluation of overall diversity, detection of invasive and/or BLAHÓ M., EGRI Á., BARTA A., ANTONI G., KRISKA G. & HORVÁTH endangered or red listed species, or surveying of popula- G. 2012: How can horsefl ies be captured by solar panels? A new tion abundance of particular species (Dodds et al., 2015). concept of tabanid traps using light polarization and electricity produced by photovoltaics. — Vet. Parasitol. 189: 353–365. Concerning the overall diversity of the horse fl y fauna, the CHVÁLA M. 1988: Family Tabanidae. In Soós Á. (ed.): Catalogue trap types used in this study caught less than half (10 out of Palaearctic Diptera. Akadémiai Kiadó, Budapest, pp. 97– of 23 species, 43%) of the possible species richness during 171. the single study season (72 trap days for each trap type). CHVÁLA M., LYNEBORG L. & MOUCHA J. 1972: The Horse However, all three trap types sampled the three most abun- of Europe (Diptera, Tabanidae). Entomological Society of Co- dant species in accordance with previous studies (Krčmar, penhagen, Copenhagen, 500 pp. 2005; Krčmar et al., 2009; Mikuška, 2010). Evaluating this CHVÁLA M., JEŽEK J. & NILSSON A. 1997: Diptera Tabanidae, result on the faunal level, species not present in the catch Horse Flies. In Nilsson A. (ed.): The Aquatic Insects of North were rare species with less than 0.31% occurrence during Europe. Apollo Boks, Stenstrup, pp. 295–309. DESQUESNES M., BITEAU-COROLLER F., BOUYE R J., DIA M.L. & FOIL the 2004–2009 period (Mikuška, 2010). According to the L. 2009: Development of a mathematical model for mechani- Shannon-Wiener diversity index, this study demonstrated

535 Mikuška et al., Eur. J. Entomol. 113: 531–536, 2016 doi: 10.14411/eje.2016.072

cal transmission of trypanosomes and other pathogens of cattle KRČMAR S., MIKUŠKA A. & RADOLIĆ V. 2010: Comparison of sam- transmitted by tabanids. — Int. J. Parasitol. 39: 333–346. pling tabanids (Diptera: Tabanidae) by four different potential DODDS K.J., ALLISON J.D., MILLER D.R., HANAVAN R.P. & SWEENEY attractants. — J. Appl. Entomol. 134: 608–613. J. 2015: Considering species richness and rarity when select- KRČMAR S., RADOLIĆ V., LAJOŠ P. & LUKAČEVIĆ I. 2014: Ef fi ciency ing optimal survey traps: comparisons of semiochemical baited of colored modifi ed box traps for sampling of tabanids. — fl ight intercept traps for Cerambycidae in eastern North Amer- Para site 21: 67. ica. — Agric. For. Entomol. 17: 36–47. KREBS S.J. 1989. Ecological Methodology. Harper Collins, New EGRI Á., BLAHÓ M., SÁNDOR A., KRISKA G., G YURKOVSZKY M., York, 652 pp. FARKAS R. & HORVÁTH G. 2012: New kind of polarotaxis gov- MAZOKHIN-PORSHNYAKOV G.A., CHERKOSOV A.D., BURAKOVA A. & erned by degree of polarization: attraction of tabanid fl ies to VISCHNEVSKAYA T.M. 1975: On the color vision of Tabanidae differently polarizing host animals and water surfaces. — (Diptera). — Zool. Zh. 54: 574–576. Naturwissenschaften 99: 407–416. MIHOK S., CARLSON D., KRAFSUR E. & FOIL L. 2006: Performance FOIL L. & HOGSETTE J. 1994: Biology and control of tabanids, sta- of the Nzi and other traps for biting fl ies in North America. — ble fl ies and horn fl ies. — Rev. Sci. Tech. OIE 13: 1125–1158. Bull. Entomol. Res. 96: 387–397. GIBSON G. & TORR S. 1999: Visual and olfacto ry responses of MIHOK S. 2007: Nzi Trap Website. URL: http://www.nzitrap. com haematophagous Diptera to host stimuli. — Med. Vet. Entomol. (last accesed 15 Apr. 2011). 13: 2–23. MIHOK S. & CARLSON D. 2007: Performance of painted plywood HALL M.J., FARKAS R. & CHAINEY J.E. 1998: Use of odour-baited and cloth Nzi traps relative to Manitoba and Greenhead traps sticky boards to trap tabanid fl ies and investigate repellents. — for tabanids and stabile fl ies. — J. Econ. Entomol. 100: 613– Med. Vet. Entomol. 12: 241–245. 618. HANSENS E. & RACE S. 2011: Rutgers, School of Environmen- MIKUŠKA A. 2010: Ecology of Horse Flies (Tabanidae) in the tal and Biological Sciences. URL: http://www.rci.rutgers. Croatian Podunavlje Floodplain and Effi ciency of Attractants edu/~insects/greenheads.htm (last accesed 19 Jan. 2011). in the Control of their Abundance. PhD thesis, Faculty of Sci- HORVÁTH G., MAJER J., HORVÁTH L., SZIVÁK I. & KRISKA G. 2008: ence, University of Zagreb, Zagreb, vi + 164 pp. Ventral polarization vision in tabanids: horsefl ies and deerfl ies MUZARI M., JONES R., SKERRATT L. & DURAN T. 2010: Feeding suc- (Diptera: Tabanidae) are attracted to horizontally polarized cess and trappability of horse fl ies evaluated with electrocuting light. — Naturwissenschaften 95: 1093–1100. nets and odour-baited traps. — Vet. Parasitol. 171: 321–326. HRIBAR L., LEPRINCE D. & FOIL L. 1991: Design f or a canopy trap ROSKOV Y., KUNZE T., PAGLINAWAN L., ORRELL T., NICOLSON D., for collecting horse fl ies (Diptera: Tabanidae). — J. Am. Mosq. CULHAM A., BAILLY N., KIRK P., BOURGOIN T. & BAILLARGEON Contr. Assoc. 7: 657–659. G. 2015: Species 2000 & ITIS Catalogue of Life. URL: http:// KRČMAR S. 2005: Response of horse fl ies (Diptera, Tabanidae) catalogueofl ife.org/annual-checklist/2015 (last accesed 17 Jul. to different olfactory attractants. — Biologia (Bratislava) 60: 2015). 611–613. TAYLOR D.B. & BERKEBILE D. 2006: Comparative effi ciency of six KRČMAR S. & GALIĆ B. 2004: Statistical analysis of the feeding stable fl y (Diptera: Muscidae) traps. — J. Econ. Entomol. 99: sites for some horse fl ies (Diptera: Tabanidae) on horses in 1415–1419. Eastern Croatia. — Ekologia (Bratislava) 23: 321–327. VAN HENNEKELER K., JONES R., SKERRATT L., FITZPATRICK L., RE ID KRČMAR S. & RUPNIK POKLUKAR D. 2011: Comparison of the effi - S. & BELLIS G. 2008: A comparison of trapping methods for ciency of three types of Canopy traps on the collection of horse Tabanidae (Diptera) in North Queensland, Australia. — Med. fl ies (Diptera: Tabanidae). — Entomol. Gener. 33: 115–123. Vet. Entomol. 22: 26–31. KRČMAR S., MIKUŠKA A. & MERDIĆ E. 2006: Response of Tabani- VOTÝPKA J., RÁDROVÁ J., SKALICKÝ T., JIRKŮ M., JIRSOVÁ D., dae (Diptera) to different natural attractants. — J. Vector Ecol. MI HALCA A.D., D’AMICO G., PETRŽELKOVÁ K.J., MODRÝ D. & 31: 262–265. LUKEŠ J. 2015: A tsetse and tabanid fl y survey of African great KRČMAR S., MIKUŠKA A. & JASIKA M. 2009: Horsefl y fa una of apes habitats reveals the presence of a novel trypanosome line- three different forest communities in the Danube river fl ood- age but the absence of Trypanosoma brucei. — Int. J. Parasi- plain in Croatia (Diptera: Tabanidae). — Entomol. Gener. 32: tol. 45: 741–748. 23–34. Received June 2, 2016; revised and accepted October 7, 2016 Published online October 31, 2016

536